Research On Trust-aware Recommendation Algorithm Based On Social Network And User Influence

With the rapid development of Internet,the problem of information overload is becoming more and more serious.How to help users find the information they are interested in from the vast information of the Internet,has become an urgent need to solve the task.The traditional collaborative filtering(CF)recommendation system has been limited due to the problem of data sparsity,cold start and scalability.Therefore,trust-aware recommendation system which uses the trust relationship among users to solve these problems has been proposed.In recent years,with the increasing popularity of social networks,Internet users begin to take the initiative to publish their preferences through social networks,and to actively express their comments on an item or statement trust certain other users and their views,trust-aware recommender system will become more and more popular.This thesis presents the trust-aware recommendation algorithm based on social network and user influence(iTARS-SNUI,Social network & User Influence),and presents the concepts of the optimal trust propagation distance and user influence used by that small-worldness of trust networks and core-periphery structure,considering two factors that the small-worldness of trust networks and user influence of social networks,using the optimal trust propagation distance and user's social status to improve trust value calculation method.The next,this thesis presents trust-aware recommender system is applied to the Tmall shopping site,make full use of interactive information between users in Tmall to mining the trust between users,and combined with the optimal trust propagation distance derived from the small-worldness of trust networks,to find more trusted neighbor users through trust propagation.Finally,use the trusted neighbor users to predict rating for Tmall user,and recommended the needed items for the Tmall user from a wide range of items,increase the interaction between users and sellers of Tmall to some extent,enhance the users satisfaction with the website indirectly and improve the trading volume of the site simultaneously.Finally,Experiments on the MovieLens dataset and Netflix dataset demonstrate that the proposed approach preferably improved recommendation accuracy and recall,especially in the case of sparse data.Therefore,this article has a certain theoretical and practical signficance.